# Craft Dataset

> Crafting datasets. Same test, different inputs.

- Skill: `attac-t/craft-dataset` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add attac-t/craft-dataset`
- Raw SKILL.md: https://api.skillmd.com/api/skills/attac-t/craft-dataset/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: attac-t (https://skillmd.com/u/attac-t)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/attac-t/craft-dataset

---


# Skill: Craft Dataset

> "Same test, different inputs."

## The Standard

1. **Named Keys**: Descriptive keys make failure output readable.
2. **Separate Files**: Shared datasets live in `tests/Datasets/`.
3. **Same Assertion Semantics**: All cases must have identical assertion logic.

## The Anti-Patterns

| Don't                         | Do                     | Why               |
|-------------------------------|------------------------|-------------------|
| Anonymous arrays              | Named keys             | Readable failures |
| Different assertions per case | Explicit tests         | Clarity           |
| Cartesian explosion           | Max 2 datasets chained | Maintainability   |
| Inline for shared data        | Extract to `Datasets/` | Reuse             |

## Real-World Examples

See [examples.md](examples.md).